The number of children who are maltreated each year in the US continues to increase. Despite this alarming statistic, primary care practitioners (PCPs)are not frequent reporters of suspected child abuse (SCAN)to government child protection agencies (CPS). The reasons for this are not yet understood. This proposed research will provide the first comprehensive description and analysis of the management of SCAN in primary care practices. This study is a prospective descriptive study of practitioner management of 16,000 childhood injuries. Data will be collected in the Pediatric Research in Office Setting (PROS)Network using a protocol that is derived from one that was piloted in a regional practice-based research network (PBRN). Four hundred PROS PCP will each gather information about 40 consecutive injury visits. They will report medical diagnosis, demographic information, and their level of suspicion for child physical abuse. Follow-up telephone calls to the PCP will elicit further information about the management of these injuries. This data will be used to construct vignettes. The response of a panel of child abuse experts to these vignettes will be used to evaluate the PCP management of suspected abuse and the rate of false positive and false negative reports to CPS. The data collected will be used to address the following specific aims concerning the care of SCAN by PCPs: (1) Identify PCP related factors that affect identification of maltreatment; (2) Identify PCP related factors that affect management of any suspicious injuries; and, (3) Assess the validity of PCP management of childhood injuries. Multivariate logistic regression modeling will be used to develop a robust model of PCP decision-making concerning SCAN. This study will provide the most thorough description ever of SCAN in primary care, thus providing important data needed for both research and policy in the area of child abuse identification and management.